125 points by janesmith 2 years ago flag hide 14 comments
john_doe 2 years ago next
This is quite an interesting approach! I wonder how the training time compares to traditional methods.
machine_learning_expert 2 years ago next
@john_doe, training times can be quite substantial, but researchers are making progress on this issue.
deep_learning_fan 2 years ago next
@machine_learning_expert, I'm sure GPUs can certainly help.
techie_friend 2 years ago next
@deep_learning_fan, yes, GPUs are already being used for training in many cases. However, TPUs are starting to take over as well.
optimization_advocate 2 years ago prev next
@deep_learning_fan, besides GPUs, quantum computing is an emerging field that could also impact training times.
another_opinion 2 years ago prev next
@machine_learning_expert, or perhaps newer training algorithms can achieve better results.
statistician 2 years ago next
@curious_reader, I agree entirely. The question is whether these methods can scale to truly massive optimization problems.
super_user 2 years ago prev next
@curious_reader, these innovations have potential to impact several industries, including logistics and manufacturing.
data_enthusiast 2 years ago prev next
@john_doe, have you seen the benchmark comparisons in the paper? It's worth taking a look.
curious_reader 2 years ago next
@data_enthusiast, so I'm not the only one thinking this is intriguing? I'd love to see more comparisons with classic optimization methods.
math_modeler 2 years ago prev next
@optimization_researcher, this is a natural evolution of optimization models. It will be interesting to see this trend accelerate.
anonymous_user 2 years ago prev next
I've heard of similar applications of neural networks, but this is taking it to a whole new level.
ai_practitioner 2 years ago next
@anonymous_user, yes, and the benefits are significant. It opens up new avenues of research for optimization problems.
opimization_researcher 2 years ago prev next
@anonymous_user, indeed, and we can use the insights from these approaches to formulate even more complex optimization problems as machine learning tasks.